Field Tests
To identify the building the ball hit, we want to check the GPS signal in order to see if the ball is in the corresponding area.
To test this behaviour, we put a GPS-reciever into a ball and saved the signal every second. The result was disillusioning. Especially on locations we threw the ball very strong or far, the GPS signal went wild and was not very accurate at all. Conclusion: The reciever cannot be put inside the ball, it must be stored somewhere on the player.
Demonstration on the roof
When we finally finished our last prototype, we were ready to test the whole thing in a prepared environment. We created some small districts at our school and could then go up to the roof to test the setup. You can see the nerdy backback which contains the beagleboard, gps device, usb hub, audio and speakers, and the memory-stick.
Due to the fact that we could only build one ball/backpack setup, we had to come up for a solution which enables a quick “change player” action. By switching the USB memory stick (Which contains another player) and by restarting the beagleboard, we could achieve this.
As you can see in the video, the ball requires calibration (the values of the accelerometer seem to depend on some environmental influences) thus the throwing was not working very well (we were also scared of breaking the ball because we have not testet yet its strength and durability). But the whole System (GPS / LED / Sonic, haptical and visual feedback) worked.
These problems were later identified and corrected. Although the time ultimately ran out and we even managed to create a graphical user interface for testing some of the interactions with the ball, we couldn’t come up with a better solution for the calibration issue. Maybe for the next version.